Merge pull request #292 from delphix/projects/merge-upstream/master

Merge remote-tracking branch '6.0/stage' into 'master'

.github/scripts/install-libkdumpfile.sh

@@ -8,7 +8,7 @@
 #
 sudo apt update
 sudo apt install autoconf automake liblzo2-dev libsnappy1v5 libtool pkg-config zlib1g-dev
-sudo apt install python3.6-dev python3.7-dev python3.8-dev
+sudo apt install python3.8-dev
 
 git clone https://github.com/ptesarik/libkdumpfile.git
 

.github/workflows/main.yml

@@ -9,10 +9,10 @@ jobs:
   # Verify the build and installation of SDB.
   #
   install:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
     strategy:
       matrix:
-        python-version: [3.6, 3.7, 3.8]
+        python-version: [3.8]
     steps:
       - uses: actions/checkout@v2
       - uses: actions/setup-python@v1
@@ -31,12 +31,12 @@ jobs:
   # the "drgn" from source (there's no package currently available).
   #
   pylint:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
     steps:
       - uses: actions/checkout@v2
       - uses: actions/setup-python@v1
         with:
-          python-version: '3.6'
+          python-version: '3.8'
       - run: ./.github/scripts/install-drgn.sh
       - run: python3 -m pip install pylint pytest
       - run: pylint -d duplicate-code -d invalid-name sdb
@@ -52,10 +52,12 @@ jobs:
   # can open kdump-compressed crash dumps for the integration tests.
   #
   pytest:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
     strategy:
       matrix:
-        python-version: [3.6, 3.7, 3.8]
+        python-version: [3.8]
+    env:
+      AWS_DEFAULT_REGION: 'us-west-2'
     steps:
       - uses: actions/checkout@v2
       - uses: actions/setup-python@v1
@@ -73,12 +75,12 @@ jobs:
   # Verify "yapf" runs successfully.
   #
   yapf:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
     steps:
       - uses: actions/checkout@v2
       - uses: actions/setup-python@v1
         with:
-          python-version: '3.6'
+          python-version: '3.8'
       - run: python3 -m pip install yapf
       - run: yapf --diff --style google --recursive sdb
       - run: yapf --diff --style google --recursive tests
@@ -98,12 +100,12 @@ jobs:
   #     pytest doesn't provide stubs on typeshed.
   #
   mypy:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
     steps:
       - uses: actions/checkout@v2
       - uses: actions/setup-python@v1
         with:
-          python-version: '3.6'
+          python-version: '3.8'
       - run: ./.github/scripts/install-drgn.sh
       - run: python3 -m pip install mypy==0.730
       - run: python3 -m mypy --strict --show-error-codes -p sdb
@@ -112,7 +114,7 @@ jobs:
   # Verify that "shfmt" ran successfully against our shell scripts.
   #
   shfmt:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
     steps:
       - uses: actions/checkout@v2
       - uses: delphix/actions/shfmt@master

sdb/command.py

@@ -30,7 +30,7 @@
 
 from sdb.target import type_canonicalize_name, type_canonical_name, type_canonicalize, get_prog
 from sdb.error import CommandError, SymbolNotFoundError
-import sdb.target as target
+from sdb import target
 
 #
 # The register_command is used by the Command class when its
@@ -45,17 +45,13 @@ def register_command(name: str, class_: Type["Command"]) -> None:
     Register the specified command name and command class, such that the
     command will be available from the SDB REPL.
     """
-    # pylint: disable=global-statement
-    global all_commands
     all_commands[name] = class_
 
 
 def get_registered_commands() -> Dict[str, Type["Command"]]:
     """
     Return a dictionary of command names to command classes.
     """
-    # pylint: disable=global-statement
-    global all_commands
     return all_commands
 
 
@@ -108,11 +104,12 @@ def help(cls, name: str) -> None:
             #
             if i == 0:
                 line = line.replace('usage: ', '')
-            print("    {}".format(line))
+            print(f"    {line}")
 
         if len(cls.names) > 1:
+            aliases = ", ".join(cls.names)
             print("ALIASES")
-            print("    {}".format(", ".join(cls.names)))
+            print(f"    {aliases}")
             print()
 
         indent = "    "
@@ -168,7 +165,7 @@ def help(cls, name: str) -> None:
                     f" case it will consume no objects as input; instead it"
                     f" will locate all objects of type '{cls.output_type}',"
                     f" and emit them as output.")
-            types = list()
+            types = []
             for (_, method) in inspect.getmembers(cls, inspect.isroutine):
                 if hasattr(method, "input_typename_handled"):
                     types.append(method.input_typename_handled)
@@ -203,7 +200,7 @@ def help(cls, name: str) -> None:
             #
             for line in inspect.getdoc(  # type: ignore[union-attr]
                     cls).splitlines()[2:]:
-                print("{}".format(line))
+                print(f"{line}")
             print()
 
     #
@@ -588,12 +585,11 @@ class Walk(Command):
     def _help_message(input_type: drgn.Type = None) -> str:
         msg = ""
         if input_type is not None:
-            msg = msg + "no walker found for input of type {}\n".format(
-                input_type)
-        msg = msg + "The following types have walkers:\n"
-        msg = msg + "\t%-20s %-20s\n" % ("WALKER", "TYPE")
+            msg += f"no walker found for input of type {input_type}\n"
+        msg += "The following types have walkers:\n"
+        msg += f"\t{'WALKER':-20s} {'TYPE':-20s}\n"
         for type_, class_ in Walker.allWalkers.items():
-            msg = msg + "\t%-20s %-20s\n" % (class_.names[0], type_)
+            msg += f"\t{class_.names[0]:-20s} {type_:-20s}\n"
         return msg
 
     def _call(self, objs: Iterable[drgn.Object]) -> Iterable[drgn.Object]:
@@ -645,11 +641,12 @@ def _call(self, objs: Iterable[drgn.Object]) -> Iterable[drgn.Object]:
         assert self.input_type is not None
         expected_type = type_canonicalize_name(self.input_type)
         for obj in objs:
-            if type_canonical_name(obj.type_) != expected_type:
+            canonical_type = type_canonical_name(obj.type_)
+            if canonical_type != expected_type:
                 raise CommandError(
                     self.name,
-                    'expected input of type {}, but received {}'.format(
-                        expected_type, type_canonical_name(obj.type_)))
+                    f'expected input of type {expected_type}, but received {canonical_type}'
+                )
 
             yield from self.walk(obj)
 
@@ -724,7 +721,7 @@ def caller(self, objs: Iterable[drgn.Object]) -> Iterable[drgn.Object]:
         out_type = None
         if self.output_type is not None:
             out_type = target.get_type(self.output_type)
-        baked = dict()
+        baked = {}
         for (_, method) in inspect.getmembers(self, inspect.ismethod):
             if not hasattr(method, "input_typename_handled"):
                 continue
@@ -762,8 +759,8 @@ def caller(self, objs: Iterable[drgn.Object]) -> Iterable[drgn.Object]:
                     pass
 
             # error
-            raise CommandError(
-                self.name, 'no handler for input of type {}'.format(i.type_))
+            raise CommandError(self.name,
+                               f'no handler for input of type {i.type_}')
 
     def _call(self,
               objs: Iterable[drgn.Object]) -> Optional[Iterable[drgn.Object]]:

sdb/commands/__init__.py

@@ -20,11 +20,11 @@
 import importlib
 import os
 
-for path in glob.glob("{}/*.py".format(os.path.dirname(__file__))):
+for path in glob.glob(f"{os.path.dirname(__file__)}/*.py"):
     if path != __file__:
         module = os.path.splitext(os.path.basename(path))[0]
-        importlib.import_module("sdb.commands.{}".format(module))
+        importlib.import_module(f"sdb.commands.{module}")
 
-for path in glob.glob("{}/*/__init__.py".format(os.path.dirname(__file__))):
+for path in glob.glob(f"{os.path.dirname(__file__)}/*/__init__.py"):
     module = os.path.basename(os.path.dirname(path))
-    importlib.import_module("sdb.commands.{}".format(module))
+    importlib.import_module(f"sdb.commands.{module}")

sdb/commands/exit.py

@@ -23,7 +23,6 @@
 
 
 class Exit(sdb.Command):
-
     "Exit the application"
 
     names = ["exit", "quit"]

sdb/commands/filter.py

@@ -113,8 +113,8 @@ def _call_one(self, obj: drgn.Object) -> Iterable[drgn.Object]:
             if not isinstance(lhs, drgn.Object):
                 raise sdb.CommandInvalidInputError(
                     self.name,
-                    "left hand side has unsupported type ({})".format(
-                        type(lhs).__name__))
+                    f"left hand side has unsupported type ({type(lhs).__name__})"
+                )
 
             if isinstance(rhs, str):
                 lhs = lhs.string_().decode("utf-8")
@@ -127,10 +127,10 @@ def _call_one(self, obj: drgn.Object) -> Iterable[drgn.Object]:
             else:
                 raise sdb.CommandInvalidInputError(
                     self.name,
-                    "right hand side has unsupported type ({})".format(
-                        type(rhs).__name__))
+                    f"right hand side has unsupported type ({type(rhs).__name__})"
+                )
 
-            if eval("lhs {} rhs".format(self.compare), {'__builtins__': None}, {
+            if eval(f"lhs {self.compare} rhs", {'__builtins__': None}, {
                     'lhs': lhs,
                     'rhs': rhs
             }):

sdb/commands/internal/__init__.py

@@ -20,7 +20,7 @@
 import importlib
 import os
 
-for path in glob.glob("{}/*.py".format(os.path.dirname(__file__))):
+for path in glob.glob(f"{os.path.dirname(__file__)}/*.py"):
     if path != __file__:
         module = os.path.splitext(os.path.basename(path))[0]
-        importlib.import_module("sdb.commands.internal.{}".format(module))
+        importlib.import_module(f"sdb.commands.internal.{module}")

sdb/commands/linux/__init__.py

@@ -20,11 +20,11 @@
 import importlib
 import os
 
-for path in glob.glob("{}/*.py".format(os.path.dirname(__file__))):
+for path in glob.glob(f"{os.path.dirname(__file__)}/*.py"):
     if path != __file__:
         module = os.path.splitext(os.path.basename(path))[0]
-        importlib.import_module("sdb.commands.linux.{}".format(module))
+        importlib.import_module(f"sdb.commands.linux.{module}")
 
-for path in glob.glob("{}/*/__init__.py".format(os.path.dirname(__file__))):
+for path in glob.glob(f"{os.path.dirname(__file__)}/*/__init__.py"):
     module = os.path.basename(os.path.dirname(path))
-    importlib.import_module("sdb.commands.linux.{}".format(module))
+    importlib.import_module(f"sdb.commands.linux.{module}")

sdb/commands/linux/dmesg.py

@@ -54,4 +54,4 @@ def pretty_print(self, objs: Iterable[drgn.Object]) -> None:
             message = drgn.cast("char *", obj) + obj.type_.type.size
             text = message.string_().decode('utf-8', 'ignore')
 
-            print("[{:5d}.{:06d}]: {:s}".format(secs, usecs, text))
+            print(f"[{secs:5d}.{usecs:06d}]: {text}")

sdb/commands/linux/internal/__init__.py

@@ -20,7 +20,7 @@
 import importlib
 import os
 
-for path in glob.glob("{}/*.py".format(os.path.dirname(__file__))):
+for path in glob.glob(f"{os.path.dirname(__file__)}/*.py"):
     if path != __file__:
         module = os.path.splitext(os.path.basename(path))[0]
-        importlib.import_module("sdb.commands.linux.internal.{}".format(module))
+        importlib.import_module(f"sdb.commands.linux.internal.{module}")

sdb/commands/linux/slabs.py

@@ -104,7 +104,7 @@ def no_input(self) -> Iterable[drgn.Object]:
         # that will be the input of the next command in the pipeline.
         #
         if self.args.s and not self.islast:
-            if self.args.s not in Slabs.FIELDS.keys():
+            if Slabs.FIELDS[self.args.s] is None:
                 raise sdb.CommandInvalidInputError(
                     self.name, f"'{self.args.s}' is not a valid field")
             yield from sorted(
@@ -154,8 +154,8 @@ def __pp_parse_args(self) -> Tuple[str, List[str], Dict[str, Any]]:
 
         for field in fields:
             if field not in self.FIELDS:
-                raise sdb.CommandError(
-                    self.name, "'{:s}' is not a valid field".format(field))
+                raise sdb.CommandError(self.name,
+                                       f"'{field}' is not a valid field")
 
         sort_field = ""
         if self.args.s:

sdb/commands/member.py

@@ -178,8 +178,7 @@ def _parse_member_tokens(
                 if not tokens[1].isdigit():
                     raise sdb.CommandError(
                         self.name,
-                        "incorrect index: '{}' is not a number".format(
-                            tokens[1]))
+                        f"incorrect index: '{tokens[1]}' is not a number")
                 is_index = True
                 idx = tokens[1]
                 sep = MemberExprSep.ARRAY
@@ -191,8 +190,7 @@ def _parse_member_tokens(
 
             if not is_index and not identifier.isidentifier():
                 raise sdb.CommandError(
-                    self.name,
-                    "{} is not an acceptable identifier".format(identifier))
+                    self.name, f"{identifier} is not an acceptable identifier")
 
             if not is_index:
                 assert idx == ""
@@ -238,8 +236,8 @@ def _validate_type_dereference(self, obj: drgn.Object,
         if kind == drgn.TypeKind.STRUCT and sep != MemberExprSep.DOT:
             raise sdb.CommandError(
                 self.name,
-                "'{}' is a struct - use the dot(.) notation for member access".
-                format(obj.type_))
+                f"'{obj.type_}' is a struct - use the dot(.) notation for member access"
+            )
 
         if kind == drgn.TypeKind.POINTER:
             assert sep in [
@@ -249,8 +247,9 @@ def _validate_type_dereference(self, obj: drgn.Object,
 
         if kind == drgn.TypeKind.ARRAY and sep != MemberExprSep.ARRAY:
             raise sdb.CommandError(
-                self.name, "'{}' is an array - cannot use '{}' notation".format(
-                    obj.type_, sep.value))
+                self.name,
+                f"'{obj.type_}' is an array - cannot use '{sep.value}' notation"
+            )
 
     def _validate_array_index(self, type_: drgn.Type, idx: int) -> None:
         """

sdb/commands/pretty_print.py

@@ -41,8 +41,7 @@ def _call(self, objs: Iterable[drgn.Object]) -> None:
 
         if handling_class is None:
             if first_obj_type is not None:
-                msg = 'could not find pretty-printer for type {}\n'.format(
-                    first_obj_type)
+                msg = f'could not find pretty-printer for type {first_obj_type}\n'
             else:
                 msg = 'could not find pretty-printer\n'
             msg += "The following types have pretty-printers:\n"

sdb/commands/pyfilter.py

@@ -46,9 +46,12 @@ def __init__(self,
             raise sdb.CommandEvalSyntaxError(self.name, err)
 
     def _call(self, objs: Iterable[drgn.Object]) -> Iterable[drgn.Object]:
-        # pylint: disable=eval-used
-        func = lambda obj: eval(self.code, {'__builtins__': None}, {'obj': obj})
+
+        def filter_cb(obj: drgn.Object) -> Optional[drgn.Object]:
+            # pylint: disable=eval-used
+            return eval(self.code, {'__builtins__': None}, {'obj': obj})
+
         try:
-            yield from filter(func, objs)
+            yield from filter(filter_cb, objs)
         except (TypeError, AttributeError) as err:
             raise sdb.CommandError(self.name, str(err))

sdb/commands/spl/__init__.py

@@ -20,7 +20,7 @@
 import importlib
 import os
 
-for path in glob.glob("{}/*.py".format(os.path.dirname(__file__))):
+for path in glob.glob(f"{os.path.dirname(__file__)}/*.py"):
     if path != __file__:
         module = os.path.splitext(os.path.basename(path))[0]
-        importlib.import_module("sdb.commands.spl.{}".format(module))
+        importlib.import_module(f"sdb.commands.spl.{module}")